University of Birmingham

Kinship in Aegean Prehistory? Ancient DNA inHuman Bones from Mainland Greece and CreteBouwman, Abigail S.; Brown, Keri A.; Brown, Terence A.; Chilvers, Elizabeth R.; Arnott,Robert; Prag, A.J.N.W.DOI:10.1017/S0068245400000265

License:None: All rights reserved

Document VersionPublisher's PDF, also known as Version of record

Citation for published version (Harvard):Bouwman, AS, Brown, KA, Brown, TA, Chilvers, ER, Arnott, R & Prag, AJNW 2009, 'Kinship in AegeanPrehistory? Ancient DNA in Human Bones from Mainland Greece and Crete', Annual of the British School atAthens, vol. 104, pp. 293-309. https://doi.org/10.1017/S0068245400000265

Link to publication on Research at Birmingham portal

Publisher Rights Statement:© The Council, British School at Athens 2009Eligibility for repository: checked July 2014

General rightsUnless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or thecopyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposespermitted by law.

•Users may freely distribute the URL that is used to identify this publication.•Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of privatestudy or non-commercial research.•User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?)•Users may not further distribute the material nor use it for the purposes of commercial gain.

Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.

When citing, please reference the published version.

Take down policyWhile the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has beenuploaded in error or has been deemed to be commercially or otherwise sensitive.

If you believe that this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we will remove access tothe work immediately and investigate.

Download date: 18. May. 2020

The Annual of the British School at Athenshttp://journals.cambridge.org/ATH

Additional services for The Annual of the British School atAthens:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

Kinship in Aegean Prehistory? Ancient DNA in Human Bonesfrom Mainland Greece and Crete

Abigail S. Bouwman, Keri A. Brown, Terence A. Brown, Elizabeth R. Chilvers, Robert Arnott and A.J.N.W.Prag

The Annual of the British School at Athens / Volume 104 / November 2009, pp 293 - 309DOI: 10.1017/S0068245400000265, Published online: 10 June 2011

Link to this article: http://journals.cambridge.org/abstract_S0068245400000265

How to cite this article:Abigail S. Bouwman, Keri A. Brown, Terence A. Brown, Elizabeth R. Chilvers, Robert Arnott and A.J.N.W.Prag (2009). Kinship in Aegean Prehistory? Ancient DNA in Human Bones from Mainland Greece andCrete. The Annual of the British School at Athens, 104, pp 293-309 doi:10.1017/S0068245400000265

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/ATH, IP address: 147.188.224.221 on 04 Aug 2014

KINSHIP IN AEGEAN PREHISTORY? ANCIENT DNA IN HUMANBONES FROM MAINLAND GREECE AND CRETE1

INTRODUCTION

In the Annual for 2000 we reported on a pilot project in which we had—successfully—searched for traces of ancient DNA (aDNA) in the skeletal remains from Grave Circle B atMycenae, as part of a cross-disciplinary project seeking to discover links of kinship betweenthe different groups of graves in which we used the methods of facial reconstruction and ofepigenetic variation as well as DNA analysis (Brown et al. 2000). We believed that such anapproach might cast light on dynastic links in a period of prehistory where there could be notexts to tell the story, and we ended that report on an optimistic note: advances in molecularbiology and the concomitant development of techniques for studying aDNA made it likelythat as well as establishing kinship relationships between burials or groups of burials (Brown2001), in conjunction with bone isotope studies (Price 1989) it would be possible to identifyindividuals who might be incomers to a particular site. This would be especially interesting ifapplied to sites such as the Grave Circles at Mycenae, where the remains are presumed to bethose of individuals who held high status in their societies, and whose family relationshipsmight reveal the underlying processes by which such status was acquired.

Broader population studies might indicate affinities between the people living in differentareas at different periods, possibly throwing light on questions such as the relationshipbetween the Minoan and Mycenaean civilizations. Ancient DNA research could also answerlong-standing questions about disease in the prehistoric Aegean, in particular to testhypotheses regarding the prevalence of malaria (Angel 1966). As well as searching directly foraDNA signatures of the malaria parasite in human bones (Sallares and Gomzi 2001), typingof globin gene mutations could determine if the skeletal indications of anaemia are indeeddue to genetic thalassaemia rather the result of dietary iron deficiency (Chilvers 2004, wherethe rationale for this approach is explained).

Progress in any of these areas is clearly dependent on the survival of aDNA in humanskeletons from sites in the eastern Mediterranean, in particular in Greece and the Aegean. Ithas become clear that the most important consideration in this regard is not thechronological age of the specimens but their thermal history, as DNA degradation occurs

1 As ever we thank Helen Clark at the BSA forcontinued advice and guidance, and especially forassistance with obtaining permits. We also thank K.A.Wardle (University of Birmingham) for bone samplesfrom Nea Nicomedia, M. Wiencke (American School ofClassical Studies, Athens) for material from Lerna, S.Chryssoulaki (Ministry of Culture, Athens) for materialfrom Karaviadena, M.Ph. Papakonstantinou (Ephorate ofPrehistoric and Classical Antiquities, Lamia) for samplesfrom Antron, L. Papazoglou-Manioudaki (NationalArchaeological Museum, Athens) for the Mycenae GraveCircle A remains, E. Palaiologou and her colleagues(Ephorate of Prehistoric and Classical Antiquities,Nafplion) for material from Mycenae Grave Circle B, W.

Cavanagh (University of Nottingham) and C.B. Mee(University of Liverpool) for bones from Kouphovouno,and Anna Lagia and Josette Renard (Universite deMontpellier) for help with the Kouphovouno material.We also thank M. Collins (York) for help with calculationof thermal ages and E.B. French for help, advice andencouragement, and Ian Barnes (University CollegeLondon) and Laura Preston (University of Cambridge)for their perceptive and helpful comments as the Annual'sreferees. The work was funded by the Institute for AegeanPrehistory, the Wellcome Trust and the Leverhulme Trust.Reports focussing on the scientific aspects of the workdescribed here have also been published (Bouwman et ai2008, Chilvers et al. 2008).

294 BOUWMAN ET AL.

more rapidly at higher temperatures. A useful measure is 'thermal age' (Smith et al. 2003),which is calculated from the temperature history of a site and its geographical location. Fromexperimental studies of DNA breakdown, together with a consideration of the oldestauthenticated detections of aDNA, it has been estimated that the limit for DNA preservationis approximately 19,000 years at io°C: hence specimens from any site that has a thermal agenormalized to io°C at > 19,000 years are unlikely to contain aDNA. When the formuladescribed in Appendix B of Smith et al. (2003) is applied, using modern data for mean annualtemperatures and assuming that while in situ specimens were not subjected to substantialseasonal fluctuations, then for the eastern Mediterranean a thermal age of ig,ooo years atio°C corresponds to approximately 3600 chronological years (Chilvers 2004). This makes itunlikely that aDNA will be present at Neolithic and Early Bronze Age sites which predate 2000BC, and suggests that the Bronze Age sites of the Minoan palace period (c. 1850-1200 BC) andof Mycenaean Greece (c.1650—1150 BC) will be at the very limits for aDNA preservation. Atthese sites we can therefore anticipate that local factors such as exposure to water and thetime that has elapsed since excavation (Pruvost et al. 2007) will be crucial in determining ifaDNA can be recovered.

This article presents the next stage in our research into aDNA: whereas the pilot wasrestricted to material from Grave Circle B, we have extended it both in time and space inorder to work with a wider corpus of samples. The sites from which we have been able to takesamples are Nea Nikomedia, Lerna, Karaviadena (Zakro), Antron (Fthiotida), Kouphovouno,and Mycenae, ranging in date from the later seventh millennium BC to the middle of thesecond and covering a geographical spread from central Macedonia to Laconia and on toeastern Crete. The timing of our work was such that through the kindness of Dr LenaPapazoglou-Manioudaki we were able in 2006 to sample the two recently rediscoveredskeletons from Shaft Grave VI at Mycenae, and through collaboration with Professor WilliamCavanagh we were also able to study material excavated between 2001 and 2002 atKouphovouno. In parallel to the DNA studies, during 2007-8 our colleague Dr ArgyroNafplioti carried out strontium isotope analyses of the human remains from both grave circlesat Mycenae (Nafplioti 2009).

Previous studies of bones and teeth from prehistoric Greece support the prediction thataDNA is recoverable from some sites dating to 2000 BC, with local factors having an importantimpact on preservation (Evison et al. 1999; Brown et al. 2000; Evison 2001). Although carriedout to the highest standards prevailing at the time, these studies took place before stringentcriteria for aDNA authenticity had been established (Cooper and Poinar 2000), and inretrospect it appears probable that some of the 'detections' of aDNA described in thesepapers were in fact due to modern contamination. Conversely, it is possible that somespecimens that gave negative results in these projects would have yielded authentic aDNA ifthe highly sensitive polymerase chain reactions (PCRs) now available had been used. As wellas continuing our work at Mycenae, we have therefore also carried out a systematic survey ofaDNA preservation at six Neolithic and Bronze Age sites in Greece and Crete, examining 8gskeletons in total, and using the most up-to-date aDNA techniques, in order to evaluate thebroader potential of aDNA research in the Eastern Mediterranean.

ANCIENT DNA

MATERIALS AND METHODS

295

1. SITES AND BONE SPECIMENS

Bone specimens are listed in TABLE 1 and the locations of the sites from which these wereobtained are shown in FIG. 1. The early Neolithic village of Nea Nikomedia is a multi-periodsettlement mound located on the central Macedonian plain, close to the south-west border ofLake Ludias marshland and excavated in 1961-4 (Rodden ig62;Wardle 1996; Angel 1973a).Calibrated radiocarbon dates for the earliest phases of occupation at the site fall into therange 6400-6000 BC (Perles 2001, 99-110). In contrast with other early Neolithic sites,burials have been found within the settlement, many in shallow, irregular pits located outsidethe houses or in the rubble of older collapsed houses rather than under the floors (Perles2001, 276-9). Lerna is located in the south-east Argolid, on the western shore of the Bay ofArgos and was occupied from the sixth to the first millennia BC; excavations were carried outbetween 1952 and 1958. During the first phases of occupation in the Neolithic it is likely thatthe site was some distance from the sea, but as sea levels continued to rise, significant regionsof the Argive plain were covered with water and Lake Lerna was formed, creating a marshyenvironment in the vicinity of the site (Zangger 1991). Although the Neolithic contextsyielded ten burials (Angel 1971, 39—41), the majority date from the Middle Bronze Age(2050/2000-1700/1675 BC). There are no burials in the Early Bronze Age contexts and laterburials are rare. The Middle Bronze Age burials are usually found either next to the houses

V •

FIG. 1. Locations of sites.

ag6

TABLE i. Bone specimens.

Site

Nea Nikomedia

Lerna

Karaviadena

AntronGrave Circle A

AntronGrave Circle B

MycenaeGrave Circle A

Date

64OO-600O BC

6th millennium BC2050-1675 BC

2OOO-17OO BC

2OOO-17OO BC

20OO-1700 BC

160O-I5OO BC

BOUWMAN ET AL.

Number*

Infant 1 (A6/1)Infant 3 (£26/1 A)Infant 5 (B6/2)NN4 (NN XLI D5/1 IV)NN13* (NNXIIIR7/1C)NN19 (NNXIXCg/id)NN5 (NN L E3 V)NN23 (NN XXIII TX-5/1)NN28 (XXVIII Ta2)Ler 220 (EN)Ler 10Ler 103Ler 81Ler125Ler 48Ler 203Ler 72Burial la (A1-A2-K3, ZK15)Burial lb (A1-A2-K3, ZK16)Burial IC (A1-A2-K3, ZK17)Burial id (A1-A2-K3, ZK18)Burial 2 (A3, ZK8)Burial 3 (A4-A7, ZK10)Burial 4-5 (K1-A-A5, ZKig)AXLVIIIbMHAXIIAIAXIIAXLVIIIbLHALIIIBIIBillBVA3.,A3.2

A3-3A4.1A4.2A4.3A4.22A4.27A5-25A5.26A6a

A6bAM

Description

Long bone fragmentLong bone fragmentLong bone fragmentLeft clavicle fragmentShort rib fragmentLong bone fragmentSecond metatarsalRight second metatarsalLeft third metatarsalMandible fragmentCranial fragmentLong bone fragmentRadius fragmentCranial fragmentPatellaCranial fragmentRadius fragmentLeft talusLeft talusLeft talusLeft talusLong bone fragmentLong bone fragmentLong bone fragmentSkull fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentMandibleMandibleMandibleMandibleMandibleMandibleMandibleMandibleMandibleMandibleClavicleClavicleUnrecorded

ANCIENT DNA 297

Site Date Number" Description

MycenaeGrave Circle B

Kouphovouno

1650-1550 BC

2OOO-17OO BC

r 5 iB52

P53H54r 5 5A56H57

r5sZ59A60A61

A62063N66N66a168A69K70A7oaA7oaiA7oa2

^ 7 o a 3KEoogKE009BKE105KE108KE171

KE173KE186KE207KE213LKE213UKE216KE220KE601KE704KE705KE706KE707KE713KE715KE716

Long bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentLong bone fragmentTibia fragmentTibia fragmentHumerus fragmentRadius fragmentRadius fragmentHumerus fragmentTibia fragmentTibia fragmentHumerus fragmentFibula fragmentRib fragmentFemur fragmentFemur fragmentSkull fragmentHumerus fragmentUlna fragmentHumerus fragmentTibia fragmentTibia fragment

According to site, inventory or museum records. For osteology reports see for Nea Nikomedia: Angel (1973a), Lagia(1993); for Lerna: Angel (1971), Lagia (1993); for Karaviadena: Arnott and Morgan-Forster (in press); Antron:A.Papathanasiou (unpublished); Mycenae: Angel (1973b), Musgrave el al. (1995); Kouphovouno; Lagia (in Cavanaghel al. 2007).

298 BOUWMAN ET AL.

or, in the case of some of the infant burials, under the floors of houses. The site ofKaraviadena is located on the eastern coast of Crete, somewhat less than a kilometre south ofthe great Middle-Late Minoan palace at Zakro. During the construction of a road from EpanoZakro to Kato Zakro a grave containing what appear to be six burials was discovered here, andsubsequently excavated in 1994 by the Greek Archaeological Service. The burials are believedto date from Middle Minoan II phases of occupation in the area (1850-1700 BC) and mayform part of a larger cemetery (at present unpublished, but see Arnott and Morgan-Forster inpress; Arnott and Stuckey 2003). Antron (Glypha Bay, Fthiotida) is located on the east coastof mainland Greece. Two grave circles (A and B) adjacent to one another were excavated in1990—1995: most of the burials were in cist graves and dated to the Middle Helladic III to LateHelladic II A periods (1750—1450 BC) (Papakonstantinou 1999a and b). Mycenae is locatedin the north-east of the Peloponnese. Grave Circles A and B date to 1675-1500 BC, GraveCircle B predating A with possibly fifty years' overlap between the two. The Grave Circlestherefore date to the very beginning of the Mycenaean age at the boundary of the Middle toLate Helladic periods. Within Grave Circle B, excavated in 1952-4, there is a developmentfrom simple cist burials to larger, deeper and richer Shaft Graves, while Grave Circle A, dugin 1876-7, comprises six Shaft Graves (Mylonas 1957). Kouphovouno, located in Laconiajustsouth of Sparta, spans the Middle Neolithic to Late Bronze Age periods (c.5000-1200 BC),and was the subject of a major excavation by the British School during 2001-2005. Twenty-seven burials were recovered, most of them from a Middle Bronze Age cemetery (2000-1700BC) and mostly from shallow earth graves (Cavanagh and Lagia, forthcoming; Cavanagh el al.2007; Lagia el al. 2007).

2 . DNA TECHNIQUES

The techniques used to extract the DNA and the regime followed in analysing the results aredescribed in detail in Bouwman el al. (2008) and Chilvers el al. (2008). Briefly, surfacecontamination, including DNA deposited on the bones by excavators and curators, wasreduced by removing the outer 1-2 mm of each bone sample with a sterile scalpel andirradiating with UV. Approximately 0.5 g was then removed from the core of each bone andany DNA present in the sample extracted by soaking the powder in a buffer. The DNA wasthen concentrated and an aliquot tested using a series of PCRs (up to 34 for each specimen)designed to amplify diagnostic regions of the mitochondrial and nuclear genomes. Themitochondrial loci that were studied were the ones containing mutations that enablemitochondrial haplogroups to be assigned, these haplogroups revealing possible maternalrelationships as the mitochondrial DNA (mtDNA) is inherited solely through the female line.The nuclear loci included many of the variable sites typed by forensic scientists in order toconstruct genetic profiles from which both maternal and paternal relationships can beinferred.

3 . ANCIENT DNA REGIME

We carried out the work in accordance with the standard criteria of authenticity for aDNAresearch (Cooper and Poinar 2000) as far as was possible. To avoid cross-contamination withDNA from previous experiments, extractions and PCRs for the Nea Nikomedia, Lerna, andKaraviadena specimens were set up in laminar flow cabinets in physically isolated labs, andthose for the Antron, Mycenae and Kouphovouno specimens were set up in similar labs but

ANCIENT DNA 299

each with an ultrafiltered air supply, with specimens handled and extractions prepared withina biological safety cabinet, and PCRs set up within a laminar flow hood. All extractions wereaccompanied by negative controls in which the entire extraction procedure was performedwithout bone material, and all PCRs were accompanied by negative controls containing waterinstead of DNA extract. Ancient DNA molecules become broken into fragments duringdiagenesis and hence are shorter than modern contaminants, and therefore the lengths ofthe template molecules in all extracts giving positive PCRs were assessed to ensure that theyfell in the anticipated range. To confirm the identity of a PCR product, the DNA was clonedbefore sequencing, as this procedure enables mixed products (e.g. specimens containingboth aDNA and modern contaminants) to be identified, and also enables aDNA sequences tobe recognised by virtue of the chemical damage they have undergone during diagenesis.Because of the small amounts of material that were available, it was not possible to carry outsome other checks that ideally would have been performed. It was not possible to performreplicate extractions for all skeletons, nor was it possible to divide the bone samples so thatportions could be sent to a second lab for independent testing, and, similarly, there wasinsufficient material to carry out tests aimed at determining the overall level of biomolecularpreservation in the specimens, such as measurements of collagen content. Corroboration ofthe human results could not be sought through study of associated animal remains, as noanimal remains were available. As mentioned above, we removed the outer 1-2 mm of eachbone prior to preparation of extracts. We have shown that even after extensive handling mostof the contaminating DNA in a bone resides in the outer 1-2 mm (Bouwman et al. 2006), andthat very little redistribution occurs if the bone is washed as in standard archaeologicalpractice (M.M. Mundee, A.S. Bowman and T.A. Brown, work in progress).

RESULTS

The results are summarized in TABLE 2. No evidence of aDNA was obtained with any of thespecimens from Nea Nicomedia, Lerna, Karaviadena, Antron Grave Circle A or MycenaeGrave Circle A. With all but one specimen from these sites, PCRs failed to give any products.The exception was sample ZK8 from Karaviadena, which gave products of the correct size withone of the mtDNA PCRs and with a PCR directed at a sex-identifying region of the nuclearDNA. However, the sequence of the mtDNA product was identical with that of E.R.Chilvers,who studied this specimen, and further examination showed that the DNA present was >425bp, longer than most genuine aDNA molecules, even those from the best-preserved material(O'Donoghue et al. 1996). These results suggest that ZK8 had become contaminated withmodern DNA from E.R.Chilvers. The possibility that aDNA was present in these bone extractsbut undetectable due to the presence of co-purifying substances that were inhibitory to PCRwas tested by 'spiking' PCRs of modern human DNA with bone extracts. These control PCRswas unaffected by addition of any bone extract, indicating that inhibitory substances wereabsent.

The results with the three specimens from Antron Grave Circle B were inconsistent butcould possibly indicate the presence of aDNA. Although mtDNA could not be detected, twoof the nuclear PCRs gave positive results with extracts of specimens BII and Bill, and a rangeof positive results were obtained after nuclear PCRs with specimen BV. Replicate PCRs didnot, however, give reproducible results and in general the yields of DNA were weak.

3°°

TABLE 2. Summary of results.

BOUWMAN ET AL.

Site PCRs attempted with each specimen8 Evidence for aDNA

Nea NicomediaLernaKaraviadenaAntron Grave Circle A

Antron Grave Circle B

Mycenae Grave Circle A

Mycenae Grave Circle B

Kouphovouno

MtD (2 ) ,MtH (2) ,GA (2)MtD (2 ) ,MtH ( 2 ) , G A ( 2 )MtD (2), MtH (2), GA (2)MtC (2), D2S11338 (2), D5S818, D10S1248,D14S1434, D16S539 (2), D18S51 (2),D22S1045, FGA (2), T H O i B (2), DYS426,M35, GA (2), MB (2)MtC (2), D2S11338 (2), D5S818, D10S1248,D14S1434, D16S539 (2), D18S51 (2),D22S1045, FGA (2), T H O i B (2), DYS426,M35, GA (2), MB (2)

MtA (2), MtG (2), MtC(2), MtF, MtD (2),MtW, MtV, CD 4 , D1S656, D2S1338,D3S1358 (2), D5S818, D6S366, D8S535,

D8S1179 (2), D10S1248, D10S2325 (2),

D14S1434, D16S539, D18S5I, D22S1045,

FGA, T H O i A (2), T H O i B, VWA (2),DYS389, DYS391, DYS393, DYS426 (2),DYS460 (2), M35, M173 (2), GA (2), MBMtA, MtG, MtC, MtF, MtD, MtW, MtV (2),CD4 (2), D1S656, D2S1338, D3S1358 (2),D5S818, D6S366, D8S535, D8S1179,D10S1248, D10S2325 (2), D14S1434,D16S539, D18S51, D22S1045, FGA,T H O i A, T H O i B, VWA (2), DYS389,DYS391, DYS393, DYS426, DYS460 (2),M35, M173, GA, MB

MtC (3), D2S1338, D10S1248, D14S1434,D16S539, D18S51, D22S1045, FGA,T H O i B, G A ( 3 ) , M B

NoneNoneNoneNone

No evidence of mtDNA.Inconsistent results fornuclear DNA in allthree bones studied.None

mtDNA in F55, F^Z59 and A62.

mtDNA and/or nuclearDNA in 7 bones.

" Numbers in brackets indicate PCRs that were carried out more than once with each specimen.

Nuclear DNA was occasionally detected in specimens from Mycenae Grave Circle B, but toosporadically for the results to be authenticated. With PCRs directed at mtDNA, 18 of the 22samples never gave a PCR product of the correct size, or if they did then that product wasconsidered to be non-endogenous to the sample because it was accompanied bycontaminated negative controls, was entirely made up of sequences containing an unusualmutation possessed by A.S. Bouwman, who performed all these extractions and PCRs (weassumed that every sequence containing this mutation was a contaminant derived from A.S.Bouwman), or was not human mtDNA. The other four samples ^ 5 5 , 1̂ 58, Z59, and A62)gave sequences which were considered to derive, at least in part, from ancient DNA. Theseresults are described in detail in the final section of this paper.

The bones we studied from Kouphovouno were excavated during 2001-2 under conditionsdesigned to minimize contamination with modern DNA, and the excavator and A.S.

ANCIENT DNA 301

Bouwman were the only people who handled these bones prior to transfer of samples to thehigh-containment laboratory at Manchester. The mitochondrial and nuclear DNA features ofthe excavator and A.S. Bouwman are known. It has been suggested that once all sequencesidentical to those of individuals who have handled a specimen are excluded, then anysequences that remain are likely to be genuine aDNA (Sampietro et al. 2006). On this basis,two of the Kouphovouno specimens (KE009B and KE105) contain mitochondrial aDNA, andsix (KE009B, KE173, KE207, KE601, KE706, KE715) contain nuclear aDNA.

DISCUSSION

Validation of aDNA research has been discussed extensively in the literature, with the 'criteriaof authenticity' proposed by Cooper and Poinar (2000) considered by many to be the goldstandard against which such work should be judged. Sometimes, however, these criteria aredifficult to meet because of the realities of biomolecular archaeology, in particular theproblems posed by the limited amount of material that is usually available for study. Museumcurators are, understandably, unwilling to allow destructive analysis of anything more thanvery small samples taken from human specimens, and their reluctance is likely to becomegreater with the growing debate regarding 'ownership' of human archaeological remains.The requirements within the 'criteria of authenticity' for multiple extractions and PCRs tocheck reproducibility of results, replication of extractions and PCRs in a second lab, andanalysis of specimens to assess the overall degree of biomolecular preservation, are reasonableif one is working with sufficient material but are not easy to satisfy if only a gram or so of boneis available. Recognising this problem, Gilbert et al. (2005) have recommended thatbiomolecular archaeologists take a cognitive and self-critical approach to authentication ofresults, which is what we attempt to do here.

A key component of a cognitive approach to authentication of aDNA detection is aconsideration of the age and preservation conditions of the specimens under study and thetime that has elapsed since their excavation, and an evaluation of whether these factors makeit possible for DNA to have survived. As temperature is the primary determinant of the rate ofDNA breakdown, the thermal history of a site can give an indication of the likelihood of aDNApresence in specimens, but such analyses are at best approximate due to difficulties indetermining factors such as seasonal temperature fluctuations and the precise conditions inthe microenvironment occupied by the buried specimens (Smith et al. 2003). However,assessment of the thermal history of a site gives an indication of the age beyond whichspecimens are unlikely to contain aDNA—placing a large burden of proof on researchersclaiming aDNA detections with older material—and helps identify in which specimens aDNAsurvival is possible, providing a starting point for self-criticism of results. In this context,judgment of the authenticity of results at one site is aided by information on the extent ofDNA survival at other sites within the same geographical region and hence likely to be ofsimilar thermal ages. Our main focus has been on Grave Circle B at Mycenae, whose thermalage is right on the limit for aDNA preservation. To aid in assessment of the DNA detectionsthat we made at Grave Circle B, we therefore surveyed aDNA survival at various other sites inGreece and Crete, from the Neolithic and Bronze Age, sites whose thermal histories also placethem at the very limits of expected survival time for aDNA.

We found possible evidence for aDNA at three of the eight sites that we studied. At Antron

302 BOUWMAN ET AL.

Grave Circle B we detected nuclear but not mitochondrial DNA in each of the three skeletonsthat we tested. These results were inconsistent with replicate PCRs failing to give reproducibleresults. The fact that only nuclear DNA could be detected is worrying, as mtDNA is present ina much higher copy number and hence rarely undetectable if genuine nuclear aDNA ispresent. On balance, we believe that these results are due to contamination of the bones withmodern DNA from previous PCR experiments carried out in the laboratory. If the results doindicate the presence of aDNA in these specimens, then that aDNA is very poorly preservedand unlikely to yield useful information. At Mycenae Grave Circle B we obtained evidence formitochondrial aDNA in four of the 2 2 skeletons that we studied. A full authentication of theseresults appears in the scientific report of our Mycenae study (Bouwman et al. 2008), and thearchaeological implications of the aDNA sequences as the final section of this article. AtKouphovouno we also obtained evidence for aDNA that, subject to more detailed assessment,we believe to be genuine because the genetic features of the aDNA differ from those of theonly two individuals who could have contaminated the bone samples.

Equally important are the negative results that we obtained. We have no evidencewhatsoever of aDNA in specimens from Nea Nicomedia, Lerna, Karaviadena, or MycenaeGrave Circle A. For the specimens from Nea Nicomedia and for Lerna no. 220 this result isfar from surprising because at 7000-8000 years these bones are substantially older than theexpected limit (3600 years) for aDNA survival in Greece based on calculations of thermal age(Chilvers 2004). The younger specimens from Lerna are dated to 2050-1675 BC and hencecloser to the 3600 year age limit, but the marshy conditions that have prevailed in the vicinityof Lerna for at least part of the period that these skeletons have been buried suggests arelatively high moisture content likely to promote DNA degradation. While these conditionsrendered it more likely that the ancient inhabitants of Lerna suffered from malaria (and thebone evidence suggested that anaemia was common), it vitiated the hopes of being able tofind the aDNA signatures of the malaria parasite and the globin gene mutations associatedwith genetic thalassaemia (Chilvers 2004).

The specimens from Karaviadena (2000-1700 BC) had previously been sampled in 2001 atManchester by Elizabeth Chilvers (nee Stuckey) as part of Amott's study of malaria in theprehistoric Aegean, and the negative results that we report here derive from that study(Arnott and Stuckey 2003; Arnott and Morgan-Forster in press). Both these bones and thosefrom Mycenae Grave Circle A (1600-1500 BC) are close to the thermal age limit and hencepossibly expected to show some indication of aDNA survival. However, those fromKaraviadena were poorly preserved at the time of excavation, being highly fragmented,suggesting that overall biomolecular preservation might be poor. The Mycenae Grave CircleA bones, excavated in 1876—7, have been housed in museums for 130 years, and it is now clearthat aDNA breakdown accelerates after excavation of bones (Pruvost et al. 2007): thus anyaDNA present in the Mycenae bones when Schliemann and Stamatakis discovered them willalmost certainly have degraded during the intervening decades.

We conclude that, although aDNA might be present in some skeletons from later centuriesof the Greek Bronze Age, it is not commonly present in Greek material from this period andis likely to be absent from older material. In reaching this conclusion, we used optimized PCRsystems in order to maximize our chances of detecting aDNA if it was present, but we also usedan ultraclean facility and took scrupulous care to remove surface contamination from thebone samples, to prevent cross-contamination with PCR products from previous experiments,

ANCIENT DNA 303

and to identify contamination that remained. We also confirmed that negative results werenot due to inhibition of PCRs by substances co-purifying with aDNA. We therefore believe thatall putative detections of aDNA from the Neolithic and Bronze Age periods of EasternMediterranean prehistory require convincing authentication, whether through self-criticismover results or through adherence to criteria of authenticity.

GRAVE CIRCLE B AT MYCENAE

Whether a group of skeletons buried in proximity to one another represents the members ofa single family can be a key question when human remains are excavated at archaeologicalsites of any age. Our work at Mycenae was conceived as a search for kinship among the peopleburied in Grave Circle B. The graves in this circle appear to have been laid out in three groupsin the south-east, north-west, and north-east sectors of the circle, with a fourth groupcomprising two graves (E and F) just east of the centre. They were named by the excavatorswith the letters of the Greek alphabet to distinguish them from those in Circle A which hadbeen given Roman numerals; later Angel identified the individual skeletons with Arabicnumerals in the order in which he studied them. Archaeologically one could only guess atwhat relationship, if any, the occupants of the four groups of graves might have to each other:did each group represent different families, or just different branches of the same family?Facial reconstruction had already identified three distinct facial types among the seven skullsthat could be reconstructed, which we thought might represent different family groups (Praget al. 1995): F55, F58, and A62 all had heart-shaped faces with wide-set cheekbones and eyes,and small, rather delicate features; F5i and Z5g had long faces with high foreheads, lanternjaws, and narrow features, while B52 had a large beaky face in a small head, and probablyrepresents a third type or family. Finally, £131 had something in common with both the firsttwo types. In terms of relative dating, these individuals covered the whole period of use of thegrave circle (c.1675-1550 BC): Z59 and £131 were buried early in the circle's use, B 52 was'early middle', and A 62 along with the three individuals in Grave F were all late. The layoutof the graves and the kinship connections suggested by facial reconstruction are shown inFIG. 2.

Ancient DNA has been used to study such relationships in a historic context (Gill et al.1994; Gerstenberger et al. 1999; Dudar et al. 2003; Gilbert et al. 2005), and so it wasintroduced here to test or to support the results suggested by facial reconstruction after a pilotproject to confirm the survival of DNA in the bones (Brown et al. 2000). Altogether we tested2 2 of the skeletons from Grave Circle B, including all those for which facial reconstructionshad been carried out, except for £131: in this case it was no longer possible to identify theassociated post-cranial bones in the Nauplia Museum and the skull itself was in too goodcondition to permit any intrusive sampling, so to our great regret it could not be tested forDNA. Our experiences with specimens from the other sites that we studied warned us that atbest we could expect to detect DNA in only a few of these skeletons, and this turned out to bethe case, with 18 of the Grave Circle B inhabitants giving entirely negative results. The fourother skeletons yielded evidence for mitochondrial but not nuclear DNA. These fourskeletons were F55, F58, Zsg, and A62. The fact that facial reconstructions were available foreach of these is perhaps not just a fortunate coincidence as the reconstructions had beenperformed on the best-preserved skulls, which one might expect to be from the skeletons

3°4 BOUWMAN ET AL.

FIG. 2. Plan of Grave Circle B showing kinship connections between graves based on facial reconstructions.Solid lines indicate probable links, broken lines possible links, and dotted lines tentative links.

displaying the best overall preservation and hence the greatest likelihood of containingancient DNA. The DNA sequences of two of these individuals, ^ 5 and 1̂ 58, were identical,but different from that of Z50,. With A62 the DNA was very poorly preserved, but from thelimited information that we could obtain we were able to establish that its sequence wasdifferent again, representing a third class. Why do we have confidence that these detectionsare of genuine aDNA and not modern contamination? First, there is the information gainedfrom other sites suggesting that aDNA can, under favourable circumstances, survive inmaterial from the Greek Bronze Age. Second, from the identity of the sequences we can

ANCIENT DNA 305

exclude the possibility that they derive from laboratory contamination that occurred after wetook our samples. Of course, the bones had been handled prior to our sampling andcontamination might have occurred then. Grave Circle B was excavated in 1952-4, using theprocedures current at that time and hence without precautions to prevent DNAcontamination. We established, however, that since excavation the bones had not beenextensively handled, except by Dr J.L. Angel and his two assistants, who carried out theosteological examination in 1954. We therefore surmise that extensive contamination bymultiple individuals is unlikely. The fact that three different DNA sequences were obtainedfrom the four skeletons therefore becomes important. We argue that the DNA that we haveidentified could be modern contamination, if: (i) ^ 5 and 1̂ 58 were handled by one or moreindividuals who did not handle, or at least did not contaminate, any of the other 20 specimensthat we studied; (ii) ZtjQ was handled by a different individual who did not contaminate anyof the other 21 bones; and (iii) A62 was similarly handled by another different individual whodid not contaminate any of the other 21 bones. This scenario is possible but we consider thealternative explanation, that these DNAs are ancient in origin, to be more likely (Bouwman etal. 2008).

As the tests only yielded data for the mitochondrial DNA, we could infer only the maternalrelationships and nothing through the male line. ^ 5 and 1̂ 58 could share a maternalrelationship. There seems little doubt that ^ 8 is a woman: even before the DNA evidence,Angel had noted that although she was tall and strongly built, 'browridges likewise agree withthe markedly female true pelvis (birth canal) and pelvis in showing female sex' (1973b, 381and table 1). Interestingly, although ^ 5 was identified as male by both Mylonas and Angelon the grounds of grave assemblage and the skeletal remains respectively, the first round ofDNA analyses suggested that this might possibly be a female skeleton, although this was laterrejected on the grounds that the female DNA results were less certain and the repeat testproved unsuccessful (Mylonas 1973, i. 46-7; Angel 19736, 379-80 and table 1; Brown et al.2000, 117 and table 1). Many of the graves in both circles contain multiple burials made overa period of time, and the sequence in Grave F seems to have been first an unidentifiedindividual (probably male), then ^ 8 after an interval long enough for the first skeleton tohave become completely disarticulated, and finally ^ 5 and ^ 1 (Angel 19736, 381; Mylonas1973, i. 48-g). The fact that F58's skeleton was still well articulated suggests that she wasburied only a few months before 1̂ 55 and his companion F5i; there is no evidence from anyof the other burials to suggest the use of a shroud that would have kept her skeleton togetherafter the connective tissue had decomposed.

F58 and first F55 were also close in age: Angel and Musgrave both reckon that he wasprobably around 33 and she was perhaps 36 years old at death (Angel 19736, 379-81;Musgrave in Prag et al. 1995, 132-3). Therefore not mother and son; the simplestinterpretation is that they were brother and sister, but they could equally be cousins whosemothers were sisters, or second cousins whose maternal grandmothers were sisters, and so on.It is of course possible that they are unrelated but just by chance have the same mitochondrialDNA, but as their particular DNA occurs in only approximately 5% of Europeans today, it ismuch more likely that they are related in one of the ways described.

The aDNA evidence tells us that Z59 does not have any maternal relationship with F55 orF58. He is not a full brother of F55 or F58 nor the son of F58. We cannot say anything abouthis paternal line. So, for example, he could in theory share a father with F55 and/or F58 but

306 BOUWMAN ET AL.

have a different mother, he could be the father of one or both, or he could be F55's son bysomeone other than ^ 8 , but our data tell us nothing on this score. The facial evidencesuggests a close relationship with Ft^ 1, though the two men were buried in different parts ofthe grave circle and are maybe three generations apart, but there are no DNA results to helpus. The fact that f"51 was given a relatively poor burial next to two very rich people in this lategrave may suggest a shift in family status and relationships.

The same conclusions apply for A62. He is the one for whom the DNA results are the leastsecure: according to the available DNA data he has no maternal relationship with 1̂ 55, ^ 8 ,or Z59, although he shares some facial features with 1̂ 55 and 758 such as the widely set eyesand cheekbones and the angle of cheek to chin. He was probably in his mid-twenties when hedied, a little earlier than 1̂ 55 and Z59 and buried in another part of the Grave Circle. Theages at death and burial dates are too close for him to have been their father, but it is alwayspossible that he was a paternal cousin. We have already suggested elsewhere that the centralposition of Grave T—probably the latest in the circle—suggests some kind of rapprochementor coming together of different branches of the family or of different families (Prag et al.1995, 128-9; Prag and Neave 1999, 141-2).

This may at first seem a rather thin result in the light of the effort that has been put intoit, and it is true that it illustrates the difficulty of applying this type of analysis to archaeologicalremains which have been out of the ground for a long time and in which aDNA is thereforegenerally poorly preserved and the problems caused by contamination with modern DNAmore acute. Nonetheless, we have shown that when hypotheses about kinship can beconstructed from existing evidence then the limited aDNA data obtainable from archae-ological remains can be used to test those hypotheses and advance understanding. Angelreckoned that of the 21 adults buried in Grave Circle B whose sex could be identified, 16 weremale and only 5 female, and he goes on to what one can best describe as a jeu d'esprit inspeculating about the possible fecundity of Mycenaean rulers and polygamous marriagecustoms of the period (19736, 389-90). The truth is that by the very nature of this prehistoricand preliterate period we know very little about social relations; that was after all one of thestarting-points of this project. So far we have no evidence of brother-sister marriage at thistime and place and the discovery of a close kinship between 1̂ 55 and ^ 8 does not changethat situation significantly. If this was indeed a sibling marriage then it was presumably madepossible by F58's high birth, but we are left to conjecture whether she was buried in this high-status and male-dominated grave circle because of a marital connection that was linked to herhigh birth, or because she held a position of authority by right of birth alone.

DNA analysis has thus enabled us to glimpse factors contributing to the organization of thehigher echelons of society at the beginning of the Mycenaean age. And for the archaeologicalscientist this project has pointed the way for future work: the results from Kouphovouno makeit very clear that where the samples are taken from freshly excavated bone and underconditions that allow as little contamination as possible, there is indeed much to be learnedabout the people whose story we are trying to uncover. That, surely, is a great step forward. Welike to think that the Mycenaeans—especially ^8—would have been pleased too.

Manchester Interdisciplinary Biocentre, University of Manchester ABIGAIL S. BOUWMAN

Manchester Interdisciplinary Biocentre, University of Manchester KERI A. BROWNManchester Interdisciplinary Biocentre, University of Manchester TERENCE A. BROWN

ANCIENT DNA 307

Manchester Interdisciplinary Biocentre, University of Manchester ELIZABETH R. CHILVERS

College of Medical and Dental Sciences, University of Birmingham ROBERT ARNOTT

The Manchester Museum, University of Manchester A.J.N.W. PRAG

BIBLIOGRAPHY

Angel, J.L. ig66. 'Porotic hyperostosis, anaemias, malarias, and marshes in the prehistoricEastern Mediterranean', Science, 153, 760-3.

1971. Lerna II: The People (American School of Classical Studies at Athens, Princeton).1973a. 'Early Neolithic people of Nea Nikomedia', in Schwidetzky, I. (ed.), Die Anfdnge

des Neolithikums vom Orient bis NordEuropa (Cologne), 103-12.19736. 'Human skeletons from grave circles at Mycenae', in Mylonas 1973, i (Athens),

379-97-Arnott, R. and Morgan-Forster, A. in press. 'Malaria and the Prehistory of the Aegean', Oxford

(BAR IS).and Stuckey, E. 2003. 'Finding the Hydra: the search for malaria in the Prehistoric

Aegean', in Polinger Foster, K. and Laffineur, R. (eds), Metron: Measuring the Aegean BronzeAge. Proceedings of the gth International Aegean Conference (Aegeum, 24; Liege and Austin, TX),209-15.

Bouwman, A.S. and Brown, T.A. 2005. 'The limits of biomolecular palaeopathology: ancientDNA cannot be used to study venereal syphilis', Journal of Archaeological Science, 32,691-702.

Chilvers, E.R., Brown, K.A. and Brown, T.A. 2006. 'Identification of the authenticancient DNA sequence in a human bone contaminated with modern DNA', AmericanJournal of Physical Anthropology, 131, 428-31.

Brown, K.A., Prag, A.J.N.W. and Brown, T.A. 2008. 'Kinship between burials from GraveCircle B at Mycenae revealed by ancient DNA typing', Journal of Archaeological Science, 35,2580-4.

Brown, T.A. 2001. 'Ancient DNA', in Brothwell, D.R., Pollard, A.M. (eds), Handbook ofArchaeological Sciences (Chichester), 301-11.

Brown, K.A., Flaherty, C.F., Little, L.M. and Prag, A.J.N.W. 2000. 'DNA analysis of bonesfrom Grave Circle B at Mycenae: a first report', BSA gtj, 115-19.

Cavanagh, W. and Lagia A., forthcoming. 'Burials from Kouphovouno, Sparta, Lakonia', inPhilippa-Touchais, A., Touchais, G., Voutsaki, S., and Wright, J. (eds) Mesohelladika(Athens).

Mee, C, Renard, J. 2007. 'Excavations at Kouphovouno, Laconia: Results from the 2001and 2002 Seasons', BSA 102, 11-101.

Chilvers, E.R. 2004. 'Ancient DNA and palaeopathology: malaria and Ancient Greece' (PhDthesis, University of Manchester Institute of Science and Technology).

Bouwman, A.S., Brown, K.A., Arnott R.G., Prag, A.J.N.W. and Brown, T.A. 2008. 'AncientDNA in human bones from Neolithic and Bronze Age sites in Greece and Crete', Journal ofArchaeological Science, 35, 2707-14.

Cooper, A. and Poinar, H. 2000. 'Ancient DNA: do it right or not at all', Science, 289, 1139.Evison, M.P. 2001. 'Ancient DNA in Greece: problems and prospects', Journal of

Radioanalytical Chemistry, 247, 673-8.

308 BOUWMAN ET AL.

Fieller, N.J., and Smillie, D.M. 1999. 'Ancient HLA: a preliminary survey', AncientBiomolecuks, 3, 1-28.

Gilbert, M.T.P., Bandelt, H.^J., Hofreiter, M. and Barnes, I. 2005. 'Assessing ancient DNAstudies', Trends in Ecology and Evolution, 20, 541-4.

Lagia, A. 1993. 'Differential diagnosis of the three main types of anaemia (thalassaemia, sicklecell anaemia, iron deficiency anaemia) based on macroscopic and radiographic skeletalcharacteristics' (MSc dissertation, University of Bradford).

Petroutsa E. and Manolis S. 2007. 'Health and diet during the Middle Bronze Age: thesite of Kouphovouno', in Mee, C. and RenardJ. (eds), Cooking up the Past: food and culinarypractices in the Neolithic and Bronze Age Aegean (Oxford), 313-28.

Myers,J.W., Myers, E.E. and Cadogan, G. (eds) 1992. The Aerial Atlas of Ancient Crete (London).Mylonas, G.E. 1957. Ancient Mycenae: The Capital City of Agamemnon (London).

(ed.) 1973. 'O TapucogKVKAOC;B zcbvMvKrjvcov, 2 vols. (Athens).Nafplioti, A. 2009. 'Mycenae Revisited, Part 2: strontium isotope ratio analysis of the human

remains from Grave Circle A', BSA 104, 279-91.O'Donoghue, K., Clapham, A., Evershed, R.P. and Brown, T.A. 1996. 'Remarkable

preservation of biomolecules in ancient radish seeds', Proceedings of the Royal Society ofLondon, series B, 263, 541-7.

Papakonstantinou, M.Ph. lggga. 'The Grave Circle B of Antron: preliminary report', inBetancourt, P.P., Karageorghis, V., Laffineur, R. and Niemeyer, W.-D. (eds), Meletemata,Studies in Aegean Archaeology Presented to Malcolm H. Wiener as he Enters his 65th Year(Aegaeum, 20; Liege and Austin, TX), 626-31.

19996. 'O TacpiKOc; KIJKXOC; A TT|C; Avcpcbvcu;', in Fleitpaynsva TOO A AmmazruioviKovv: HITepKpepsia TOV MvKrivaikov Koafiov (Aa/xia 1994) (Lamia), 171-80.

Papazoglou-Manioudaki, L., Nafplioti, A., Musgrave, J.H., Neave, R.A.H. and Prag, A.J.N.W.2009. 'Mycenae Revisited, Part 1: Stamatakis, Schliemann, and two new faces from ShaftGrave VI', BSA 104, 233-77.

Perles, C. 2001. The Early Neolithic in Greece (Cambridge).Prag, A.J.N.W. and Neave, R. 1999. Making Faces: Using Archaeological and Forensic Evidence, corn

repr. of 1997 edn. (London).Musgrave, J.H., Neave, R.A.H. and Thimme, D.I. 1995. 'Seven Faces from Grave Circle

B at Mycenae', BSA go, 107-36.Price, T.D. (ed.) ig8g. The Chemistry of Prehistoric Bone (New York).Pruvost, M., Schwarz, R., Bessa Correia, V., Champlot, S., Braguier, S., Morel, N., Fernandez-

Jalvo, Y, Grange, T. and Geigl, E.-M. 2007. 'Freshly excavated fossil bones are best foramplification of ancient DNA', Proceedings of the National Academy of Sciences of the United Statesof America 104, 73g~44.

Rodden, R.J. ig62. 'Excavations at the Early Neolithic Site of Nea Nikomedeia, GreekMacedonia (ig6i Season)', Proceedings of the Prehistoric Society, 28, 267-88.

Sallares, R. and Gomzi, S. 2001. 'Biomolecular archaeology of malaria', Ancient Biomolecules,3. !95-2 13-

Sampietro, M.L., Gilbert, M.T.P., Lao, O., Caralelli, D., Lari, M., Bertranpetit, J. and Lalueza-Fox, C. 2006. 'Tracking down human contamination in ancient human teeth', Molecularand Biological Evolution, 23, 1801-7.

Smith, C.I., Chamberlain, A.T., Riley, M.S., Stringer, C. and Collins, M.J. 2003. 'The thermal

ANCIENT DNA 309

history of human fossils and the likelihood of successful DNA amplification', Journal ofHuman Evolution, 45, 203-17.

Wardle, K.A. (ed.) 1996. Nea Nikomedia 1: The Excavation of an Early Neolithic Village in NorthernGreece 1961-1964 (BSA supp. vol. 25; London).

Zangger, E. 1991. 'Prehistoric coastal environments in Greece: the vanished landscapes ofDimini Bay and Lake Lerna', JFA 18, 1-15.